The main topic of this study is the mathematical modelling of bubble size distributions in an aerated stirred tank using the population balance method. The air-water system consisted of a fully baffled vessel with a diameter of 0.29 m, which was equipped with a six-bladed Rushton turbine. The secondary phase was introduced through a ring sparger situated under the impeller. Calculations were performed with the CFD software CFX 14.5. The turbulent quantities were predicted using the standard k-ε turbulence model. Coalescence and breakup of bubbles were modelled using the MUSIG method with 24 bubble size groups. For the bubble size distribution modelling, the breakup model by Luo and Svendsen (1996) typically has been used in the past. However, this breakup model was thoroughly reviewed and its practical applicability was questioned. Therefore, three different breakup models by Martínez-Bazán et al. (1999a, b), Lehr et al. (2002) and Alopaeus et al. (2002) were implemented in the CFD solver and applied to the system. The resulting Sauter mean diameters and local bubble size distributions were compared with experimental data.

The main topic of this study is the experimental measurement and mathematical modelling of global gas hold-up and bubble size distribution in an aerated stirred vessel using the population balance method. The air-water system consisted of a mixing tank of diameter T = 0.29 m, which was equipped with a six-bladed Rushton turbine. Calculations were performed with CFD software CFX 14.5. Turbulent quantities were predicted using the standard k-ε turbulence model. Coalescence and breakup of bubbles were modelled using the homogeneous MUSIG method with 24 bubble size groups. To achieve a better prediction of the turbulent quantities, simulations were performed with much finer meshes than those that have been adopted so far for bubble size distribution modelling. Several different drag coefficient correlations were implemented in the solver, and their influence on the results was studied. Turbulent drag correction to reduce the bubble slip velocity proved to be essential to achieve agreement of the simulated gas distribution with experiments. To model the disintegration of bubbles, the widely adopted breakup model by Luo & Svendsen was used. However, its applicability was questioned.